Norifumi Watanabe, Masafumi Matsuhara, G. Chakraborty, H. Mabuchi
{"title":"Method to Present Situation Aware Suitable Reply Candidate on Mobile Terminal","authors":"Norifumi Watanabe, Masafumi Matsuhara, G. Chakraborty, H. Mabuchi","doi":"10.1109/ICAwST.2019.8923371","DOIUrl":null,"url":null,"abstract":"In recent years, the number of users using a mobile terminal such as a smart phone and a small mobile terminal such as a smart watch is increasing. In addition, SNS users are increasing year by year. Therefore, necessity for creating messages on a mobile terminal through SNS is increasing. Since the operation screen is small on a mobile terminal, keys for character input are small. Therefore, creating a message on a mobile terminal takes time and typos are common. In particular, when creating a reply after receiving a message via SNS on a small mobile terminal, it may be difficult to respond quickly. Smooth communication between users may not be possible.We have proposed a method to present multiple relevant situation-aware reply candidates to the user corresponding to different situations according to the topic of the received sentence. By the proposed method, the user can select an appropriate reply without performing character input and message revision. The purpose of our proposed method is to support smooth communication between users. The system collects tweets to present a reply candidate. The collected tweets are divided into sentences. Tweets are often written in spoken language on variety of topics. Since the collected tweets may contain noise that is not appropriate as a reply, only suitable tweets are presented to the user as reply candidates by clustering and scoring them. The sentences are divided into different situations by clustering. The sentences with the highest score from each class is presented as a candidate reply. Since the proposed reply indicates different situations, it is expected that most of the time a suitable reply candidate will be presented to the user.","PeriodicalId":156538,"journal":{"name":"2019 IEEE 10th International Conference on Awareness Science and Technology (iCAST)","volume":"86 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 IEEE 10th International Conference on Awareness Science and Technology (iCAST)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICAwST.2019.8923371","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
In recent years, the number of users using a mobile terminal such as a smart phone and a small mobile terminal such as a smart watch is increasing. In addition, SNS users are increasing year by year. Therefore, necessity for creating messages on a mobile terminal through SNS is increasing. Since the operation screen is small on a mobile terminal, keys for character input are small. Therefore, creating a message on a mobile terminal takes time and typos are common. In particular, when creating a reply after receiving a message via SNS on a small mobile terminal, it may be difficult to respond quickly. Smooth communication between users may not be possible.We have proposed a method to present multiple relevant situation-aware reply candidates to the user corresponding to different situations according to the topic of the received sentence. By the proposed method, the user can select an appropriate reply without performing character input and message revision. The purpose of our proposed method is to support smooth communication between users. The system collects tweets to present a reply candidate. The collected tweets are divided into sentences. Tweets are often written in spoken language on variety of topics. Since the collected tweets may contain noise that is not appropriate as a reply, only suitable tweets are presented to the user as reply candidates by clustering and scoring them. The sentences are divided into different situations by clustering. The sentences with the highest score from each class is presented as a candidate reply. Since the proposed reply indicates different situations, it is expected that most of the time a suitable reply candidate will be presented to the user.